www.nature.com/scientificreports

OPEN Colonization kinetics and implantation follow‑up of the sewage microbiome in an urban wastewater treatment plant Loïc Morin1, Anne Goubet2, Céline Madigou2, Jean‑Jacques Pernelle2, Karima Palmier1, Karine Labadie3, Arnaud Lemainque3, Ophélie Michot4, Lucie Astoul4, Paul Barbier5, Jean‑Luc Almayrac4 & Abdelghani Sghir5*

The Seine-Morée wastewater treatment plant (SM_WWTP), with a capacity of 100,000 population- equivalents, was fed with raw domestic wastewater during all of its start-up phase. Its microbiome resulted from the spontaneous evolution of wastewater-borne microorganisms. This rare opportunity allowed us to analyze the sequential microbiota colonization and implantation follow up during the start-up phase of this WWTP by means of regular sampling carried out over 8 months until the establishment of a stable and functional ecosystem. During the study, biological nitrifcation– denitrifcation and dephosphatation occurred 68 days after the start-up of the WWTP, followed by focs decantation 91 days later. High throughput sequencing of 18S and 16S rRNA genes was performed using Illumina’s MiSeq and PGM Ion Torrent platforms respectively, generating 584,647 16S and 521,031 18S high-quality sequence rDNA reads. Analyses of 16S and 18S rDNA datasets show three colonization phases occurring concomitantly with nitrifcation, dephosphatation and foc development processes. Thus, we could defne three microbiota profles that sequentially colonized the SM_WWTP: the early colonizers, the late colonizers and the continuous spectrum population. Shannon and inverse Simpson diversity indices indicate that the highest microbiota diversity was reached at days 133 and 82 for prokaryotes and eukaryotes respectively; after that, the structure and complexity of the wastewater microbiome reached its functional stability. This study demonstrates that physicochemical parameters and microbial metabolic interactions are the main forces shaping microbial community structure, gradually building up and maintaining a functionally stable microbial ecosystem.

Te wastewater treatment process is based on the use of sludge microbial populations to treat domestic and indus- trial pollutants. Tese populations constitute a complex ecosystem with biomass concentration approximating 2–10 g L−11, with the majority aggregated into structures called focs. Te foc structure represents a protection strategy for microorganisms against predation as well as toxic chemicals, meanwhile allowing efcient uptake of nutrients. Tese focs may contain up to ­1010 prokaryotes mL­ −1 and 10­ 6 micro-eukaryotes ­mL−1. Molecular approaches reveal that they ofen share persistent prokaryotic and eukaryotic core species stably retained over time, including among others, members of the Proteobacteria, Bacteroidetes, Firmicutes, and Actinobacteria phyla­ 2–4. However, in comparison with prokaryotes, micro-eukaryotic diversity has benefted relatively little from modern molecular tools and high throughput sequencing technologies. Te few sequencing-based analyses

1Institut de Biologie Intégrative de la Cellule, Université Paris Saclay, 91405 Orsay Cedex, France. 2INRAE, PROSE, Université Paris-Saclay, 92761 Antony, France. 3Genoscope, Institut de Biologie François‑Jacob, CEA, Université Paris-Saclay, 91057 Evry, France. 4Laboratoire SIAAP Site Seine Amont, Usine Marne Aval, 100 rue de la Plaine, 93160 Noisy‑Le‑Grand, France. 5Génomique métabolique, Genoscope, Institut de Biologie François Jacob, CEA, CNRS, Université d’Evry, Université Paris-Saclay, 91057 Evry, France. *email: [email protected]

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 1 Vol.:(0123456789) www.nature.com/scientificreports/

performed on samples from WWTP or sewage treatment bioreactors suggest that large knowledge gaps in diversity and functional ecology of this group of microorganisms­ 5–8 exist, preventing accurate defnitions of their identity, diversity and their roles in the depollution process. To provide a holistic view of the functioning of whole ecosystems, major fundamental studies are necessary for assessing the network of interactions between all kinds of microbes (Bacteria, Archaea, Eukarya and viruses) and with their environment. Tis will permit the inference of the ecological rules guiding assembly of complex microbial communities and their functional impli- cations across temporal gradients, and biological and physicochemical perturbations, which still await discovery. Such studies should help anticipate the structure and activity of microbial communities and consequently the functional stability of the ecosystem. Recent studies have reported temporal variations in both composition and structure of microbial communities­ 9,10. To the best of our knowledge, no work has studied the colonization kinetics and the estab- lishment of the wastewater microbiome through to the constitution of a complex and functional microbial ecosystem. In the present study, we are taking advantage of the start-up of a domestic WWTP to achieve full characterization of 23 time series samples from the SM_WWTP over 236 days. We hence, followed the coloni- zation kinetics of wastewater-borne microorganisms from March 3rd through October 31st, 2014, using high throughput prokaryotic 16S and eukaryotic 18S rRNA gene sequencing, until the establishment of a complex functional and stable ecosystem. Results Physicochemical and overall molecular diversity analyses of microbial populations. Te infor- mation regarding plant localization and process description, variation of plant operational parameters and phys- icochemical conditions, treatment performance, are detailed in Supplementary material online (Fig. S1, Fig. S2 and Table S1). Nitrogen measurement defnes three periods: (1) Te frst period lasts for 11 weeks (13–40 days) during which ammonia in the aerobic basin is at its maximal concentration. (2) In the second period that lasts about 1 month, the ammonia starts decreasing concomitantly with the appearance of nitrite (day 40) and nitrate (day 68), until day 133 (Fig. 1A). (3) In the third period (133–236 days) nitrite is completely oxidized to nitrate at day 133 and remains relatively stable during the rest of the time. Biological dephosphatation started at day 40, and phosphorus concentration fuctuated from day 68 through day 133, to be completely reduced over this third period (133–236 days). Flocs decantation appeared 91 days afer the start-up of the SM_WWTP. Principal component analyses of the microbiota 16S and 18S rDNA data sets, revealed a diversifcation in microbial composition between samples, concomitantly occurring with the physicochemical transformations (e.g. nitrifcation, dephosphatation and foc development processes). Tis analysis indicates the constitution of three main homogeneous prokaryotic groups (Fig. 1B) whereas 18S rDNA tags indicates four eukaryotic groups (Fig. 1C). Afer stabilization, the plant could efectively remove 97.8% of chemical oxygen demand (COD) and 99.6% biological oxygen demand (BOD) from the sewage (Table S1). A total of 584,647 and 521,031 high quality reads were generated from 23 SM_WWTP rDNA amplicon sequencing (Table 1), with an average of 25,419 and 22,653 reads per sample for bacteria and eukaryotes respec- tively. Phylotype richness calculation for individual samples based on the construction of rarefaction curves shows a complete saturation of microbial diversity (Fig. S3A, Fig. S3B). Sequence read clustering, based on 97% sequence similarity reveals, the occurrence of 6,696 bacterial operational taxonomic units (OTUs) and 1,423 eukaryotic OTUs (Table 1). Taxonomic afliations using Silva database-132 shows that the 6,696 bacterial OTUs afliated with 36 phyla among which we describe 30 persistent OTUs, occur at various abundancies through- out the study (Table 1). Among the 1,423 eukaryotic OTUs, we describe 19 persistent OTUs afliating with 15 phyla (Table 1). Shannon and Inverse Simpson diversity indices show the occurrence of two periods, the frst one where microbiota increase in numbers and diversity and the second period where the microbiota reaches its steady state (Fig. 1F,G).

Overall description of wastewater microbiota phylogenetic groups. Te most predominant phy- logenetic groups within the Bacteria domain are the Gram-negative bacteria, represented by Proteobacteria and Bacteroidetes, and the Gram-positives represented by Firmicutes and Actinobacteria, averaging 64.8% and 13.1% of the total number of OTUs respectively. Planctomycetes, Chlorofexi, Verrucomicrobia, and Acidobacteria, rep- resent an average of 17.6% of the total OTUs. Te remaining 28 minor phylogenetic groups account for only 4.5%. In terms of abundance, nine phylogenetic groups, Proteobacteria, Bacteroidetes, Planctomycetes, Chloro- fexi, Firmicutes, Actinobacteria, Acidobacteria, Gemmatimonadetes, and Patescibacteria (Saccharibacteria) rep- resent up to 97.6% of the total reads (Fig. S5A), whereas the remaining 27 minor phyla made up only 2.4% of the total reads. Inside these phyla, the distribution of subphyla follows the same pattern, with major and minor sub- phyla (Fig. S5A). On the other hand, the distribution of genera abundance is characterized by a large diversity, i.e. many genera with a low abundance. However, some genera emerge with a slightly more elevated abundance; such is the case for Acinetobacter, an unknown genus from Gammaproteobacteria, Terrimonas, and an unknown genus from Saprospiraceae, and Flavobacterium, that together represent 20.6% of total reads (Fig. S5B). Fifeen phyla representing the Eukarya domain colonized the SM_WWTP. Te most predominant in terms of percent of OTUs are Nucletmycea, Holozoa, Amoebozoa, Rhizaria, Alveolata (Ciliophora, Apicomplexa), Strameno- piles, Discoba, Chloroplastida, and Protalveolata, representing an average of 89% of the total OTUs. Unknown and Multi-afliation phyla make up 10% of the total OTUs. Te remaining six minor phyla afliate with Apu- sozoa, Rhodophycea, Cryptista, Haptista, Metamonada, and Dinofagellata. Altogether, they totalize 1% of the OTUs. In terms of abundance, seven predominant phyla, Holozoa, Rhizaria, Nucletmycea, Stramenopiles, Cili- ophora, Discoba, and Amoebozoa, represent > 98.77% of the total 18S rDNA reads. Te remaining ten eukaryotic phyla made up only 1.23% of the total eukaryotic reads (Fig. S6A). Inside these seven predominant phyla, the

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 2 Vol:.(1234567890) www.nature.com/scientificreports/

AB

60 8 55.6 Effluent_N-NH4 Effluent_N-NO2

7 )

50 Effluent_N-NO3 ) 65 % -1 ) 9.

-1 Effluent_P-PO4

6 (1 gL L 5.2 Effluent_Ptotal 2 (m

40 PC mg 5 3( 30 4 4.1 PC1 (31.99%) 3 20 P-PO4_Ptotal C

NH4_NO2_NO 2

10 ) 1 15.13%

0 0 2( PC 3 4 2 9 6 3 0 7 4 1 6

18 21 28 29 30 31 32 33 34 35 40 68 82 Days 13 140 147 15 18 18 19 20 21 21 22 23 23

Start of the Flocs appearance nitrification process (D91) PC1 (20.26%) (D68) DE TU s OTUs cO ti otic yo Eukary Procar e e anc d Abundanc Abun 1 1 2 2 2 2 2 2 6 8 1 1 1 1 1 20 21 21 22 231 2 2 2 3 3 3 4 40 68 82 1 14 14 15 1 189 1 1 1 13 16 21 28 3 33 3 96 10 17 24 8 33 40 47 8 89 0 31 36 36 1 0 5 0 2 54 3 8 9 3 6 8 3 0 5 2 3 6 3 0 7 4 3 0 7 4 2 Days Days

FG 4,64 4,80 6,00 120 60 5,45 4,30 5,50 100 50 x x e de nd n ndex ndex 3,80 5,00 80 40 i y yi yi yi t 3,30 4,50 i 60 30 rs rsit ve 2,80 4,00 40 20 ve Di Diversit Di Diversit 2,30 3,50 20 10 1,80 3,00 0 0 Days Days

Figure 1. (A) Te aerobic basin physicochemical parameter evolution through the 236 days. Blue scale color for P-PO4 and total phosphorus concentration only. Colored lines delimit the three periods of physicochemical parameter evolution: Pink represents the frst period, blue, the intermediate period, and green represents the third period. (B) PCoA analysis of the 16S rDNA tags of the time series samples. Te X and Y-axis explained 51.4% of the correlation between samples. PCoA separated the time series samples into three distinct groups constituting 3 phases. Phase 1, blue boxes (day 13–day 40); phase 2, blue and red boxes (days 82 and 68); and phase 3, red boxes (day 133–day 236). (C) PCoA analysis of the 18S rDNA tags of the SM_WWTP time series samples. Te X and Y-axis explained 35.33% of the correlation between samples. (D,E) Hierarchically clustered heat map fngerprinting of the top 100 microbial OTUs respectively for bacteria (D), and eukaryotes (E), of the 23 time-series SM_WWTP samples. (F,G) Evolution of ecological diversity indices within the 23 time-series SM_WWTP. Scale colors: Red for Eukarya and blue for Bacteria. (F) Evolution of Shannon diversity indices within the SM_WWTP. (G) Evolution of Inverse Simpson diversity indices within the time series SM_WWTP. Scale colors: Red for Eukarya and blue for Bacteria.

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 3 Vol.:(0123456789) www.nature.com/scientificreports/

Domain Number of reads Number of OTUs Archaea 306 18 Bacteria 584,647 6,696 Number of Phyla 36 Persistent ­OTUsa 30 Eukarya 521,031 1,423 Number of Phyla 15 Persistent ­OTUsa 19

Table 1. Summary of the SM_WWTP microbiome 16S/18S rRNA gene sequences and their afliation at the phylogenetic domain level. a Persistent OTU: Any OTU who’s number of reads ≥ 1 sequence found in all samples.

subphyla follow the same distribution, with major and minor subphyla (Fig. S6A). In contrast to prokaryote genera distribution, eukaryotes are characterized by seven predominant genera: Rhogostoma, Phascolodon, Mal- lomonas, Petalomonas, an unknown genus from Rhinosporideaceae, an unknown genus from Pseudoscorpiones, and an unknown genus from Adinetida, that represents 56.65% of the total reads, not including Cryptomycota, a poorly known phylogenetic group (Fig. S6B).

Colonization kinetics and phylogenetic diversity of microbial populations. Overall, temporal changes in the phylogenetic composition and abundancy of OTUs show a continuous increase during the frst period. Te number of eukaryotic and bacterial phyla reached their maximum at day 140, and 154 respectively (Fig. S4). However, Shannon’s and inverse Simpson diversity indices show that Eukarya and Bacteria reached their greatest diversity between day 82 and day133 respectively (Fig. 1F,G). We show a net growth dependency between bacteria and eukaryotes, as is shown by the respective maximum and minimum percent of total phyla and OTU fuctuation in the time series samples (Fig. S4). If we take into account the evolution of physico- chemical parameters (Fig. 1A), PCoA (Fig. 1B), and heat maps (Fig. 1D), all of them indicate a clear distinction between three bacteria profles. Te period from day 13 through day 40 represents the frst profle; the inter- mediate profle is represented by the period from day 40 through 133; and the third profle is represented by the period from day 133 through day 236. However, PCoA and heat maps analyses show the distinction of four eukaryotic profles (Fig. 1C,E). Figure 1D,E, illustrate the microbial community profle at OTU level in order to better assess the diferences in microbial community composition in the SM_WWTP time series samples. As indicated in both fgures, remarkable dissimilarities at both microbial composition and relative abundances were observed. Canonical correspondence analysis (CCA) of the relation between bacterial community compositions and physicochemical properties of the time series wastewater samples showed that physicochemical properties such as COD, BOD, PO4, N-NH4, NTK correlated to the frst phase predominant bacterial communities while nitrite and nitrate remarkably correlate to the second and the third phase bacterial communities respectively (Fig. S4B).

Bacterial colonization kinetics at the phylum level. Troughout the study, the Gram-negative bac- teria, Proteobacteria and Bacteroidetes, made up an average of 79.1% [59–99%] of total 16S rDNA sequences and to lesser extent the Gram-positive bacteria; Firmicutes and Actinobacteria represent an average of only 4.9% [1–16%] (Fig. 2B,C). However, the four phyla represent the most predominant prokaryotic populations present during the frst and the second period of colonization (Fig. 2A), where Proteobacteria and Bacteroidetes together made up between 93 and 99% of the total bacterial sequences while the Gram-positives made up between 1 and 5%. Gram-negative and Gram-positive population abundance evolution throughout the study is quite diferent. While the Proteobacteria and Bacteroidetes evolve in an antagonistic way, Firmicutes and Actinobacteria evolve simultaneously and show the same growth trends (Fig. 2B,C). Te appearance of a radiation of bacterial phyla afliated with Planctomycetes, Chlorofexi, Acidobacteria, Gemmatimonadetes, Patescibacteria, Nitrospira, Arma- timonadetes, (Fig. 2D,E) occurs over the second period, (starting from day 35 through day 133). Teir appear- ance represents the beginning of the second phase of colonization.

Bacterial colonization kinetics at the order level. Within the Proteobacteria phylum, two waves of bacterial phylogenetic groups are distinguished. Te frst wave of proteobacterial colonizers is composed of nine orders: Pseudomonadales, Caulobacterales, Aeromonadales, Rhodobacterales Sphingobacterales, Rhizobiales, Burkholderiales, Enterobacteriales, and Bdellovibrionales. Tey reach their maximal growth rate during the frst period of colonization, but their abundance decreases to a minimum level over the second period. Tey maintain their presence at a very low population level or disappear from the WWTP (Fig. 2F). Nitrosomonadales, Myxo- bacteriales, Competibacterales, Xanthomonadales, and Dongiales represent the second wave (Fig. 2G). Within the Bacteroidetes–Cytophaga–Flavobacter group, Bacteroidales, Cytophagales, and Flavobacteriales orders represent the frst wave of colonizers over the frst period (Fig. 2H). Sphingobacteriales, Chitinophagales and two minor orders, Ignavibacteriales and Latescibacteriales represent the second wave of colonizers (Fig. 2I). For both the Proteobacteria and Bacteroidetes orders, the second wave of colonizers appears 40 days afer the start-up of the SM_WWTP.

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 4 Vol:.(1234567890) www.nature.com/scientificreports/

A

100%

80%

60%

Reads 40%

20%

0% 6 3 0 7 4 6 3 0 7 4 2 9 1 28 30 33 35 40 68 82 21 13 16 Days 23 19 20 21 21 22 23 13 14 14 15 18 18

Proteobacteria Bacteroidetes Planctomycetes Chloroflexi Firmicutes ActinobacteriaAcidobacteriaGemmatimonadetesPatescibacteriaNitrospirae Armatimonadetes Verrucomicrobia Hydrogenedentes Latescibacteria WPS-2 Fusobacteria Epsilonbacteraeota FibrobacteresSynergistetes Tenericutes Cyanobacteria BRC1 RsaHF231 Dependentiae Elusimicrobia OmnitrophicaeotaSpirochaetesDadabacteriaKiritimatiellaeota Deinococcus-Thermus Margulisbacteria FCPU426 CloacimonetesZixibacteriaNitrospinae Multi-affiliation B C 10% ActinobacteriaFirmicutes 10% 100% ProteobacteriaBacteroidetes 50% 8% 8% 80% 40% 60% 30% 6% 6% 40% 20% 4% 4% Reads Reads Reads Reads 20% 10% 2% 2% 0% 0% 0% 0% 3 2 6 0 4 6 7 3 7 2 6 0 4 6 13 21 30 35 68 13 21 30 35 68 Days Days 14 13 18 19 21 22 23 13 14 18 19 21 22 23

DE

Planctomycetes Chloroflexi NitrospiraeArmatimonadetes 16% AcidobacteriaGemmatimonadetes 5% Hydrogenedentes WPS-2 14% 4% 12% Patescibacteria 10% s 3% 8% 2%

Reads 6% Read 4% 1% 2% 0% 0% 3 0 7 4 2 9 6 3 0 7 4 1 6 3 7 2 6 0 4 6 13 16 21 28 30 33 35 40 68 82 13 21 30 35 68 Days Days 13 14 14 15 18 18 19 20 21 21 22 23 23 13 14 18 19 21 22 23 FG Pseudomonadales Caulobacterales Aeromonadales Rhodobacterales Nitrosomonadales Myxococcales Sphingomonadales Rhizobiales Competibacterales Xanthomonadales Enterobacteriales Bdellovibrionales 40% 2,00% 10% Dongiales 35% 30% 1,50% 8% s 25% 6% 20% 1,00% Reads Read 15% Reads 4% 10% 0,50% 5% 2% 0% 0,00% 0% 3 7 2 6 0 4 6 6 3 7 2 6 0 4 30 35 68 13 21 Days 13 21 30 35 68 Days 13 14 18 19 21 22 23 23 HI13 14 18 19 21 22 10% 30% 12% 0,8% Bacteroidales Sphingobacteriales 25% 10% 8% Chitinophagales 0,6% s s s Cytophagales s 20% 8% 6% Flavobacterales Ignavibacteria 15% 6% 0,4% Read Read Read Read Latescibacteria 4% 10% 4% 0,2% 2% 5% 2% 0% 0% 0% 0,0% 2 6 0 4 6 7 3 3 7 2 6 0 4 6 13 21 30 35 68 Days 13 21 30 35 68 14 18 19 21 22 23 13 13 14 18 19 21 22 23 Days

Figure 2. Relative abundance expressed as percent of total 16S rDNA reads and colonization kinetics of the main bacterial phyla over 236 days of the SM_WWTP colonization. (A) Relative abundance (%) of bacteria at the phylum taxonomic level within the 23 time-series SM_WWTP samples. Nine phylogenetic groups (in bold) represent up to 97.6% of the total reads. Colored lines delimit the three periods of bacterial evolution. Pink represents the frst period, blue, the intermediate period, and green represents the third period. (B) Colonization kinetics of Proteobacteria and Bacteroidetes. Scale color: blue for Proteobacteria; red for Bacteroidetes. (C) Colonization kinetics of Actinobacteria and Firmicutes. Scale color: blue for Firmicutes; red for Actinobacteria. (D–G) Colonization kinetics of the remaining major bacterial phyla representing the late colonizers (D,E), and the frst and late Proteobacteria colonizers at the phylogentic order level, in (F,G). (H,I) Bacteroidetes colonization kinetics at the phylogenetic order level over the 236 days of microbiota evolution within the SM_WWTP. Te frst colonizers (H) and late colonizers (I). Y-axis scale color: Green, for Flavobacteriales only, pink, for Latescibacteria only.

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 5 Vol.:(0123456789) www.nature.com/scientificreports/

A 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 13 16 21 28 30 33 35 40 68 82 133140 147154 182189 196203 210217 224231 236 ThermomonasStenotrophomonas Dokdonella PseudomonasAcinetobacter Unknown family (Pasteurella) Multi-affiliation GKS98) Aquabacterium Acidovorax Pseudochrobactrum Phenylobacterium Brevundimonas Unknown family (Saccharimonadales) PedobacterFlavobacterium Dyadobacter Unknown genus (OLB8-1) Terrimonas Ferruginibacter B 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 13 16 21 28 30 33 35 40 68 82 133140 147154 182189 196203 210217 224231 236 Multi-affiliation (Eukaryota) Unknown genus (Rhizophydiaceae) Unknown class (Cryptomycota) Unknown family (Rhinosporideaceae)) Unknown family (Pseudoscorpiones) Unknown family (Adinetida) Protosporangium Rhogostoma Unknown class (Oomycetes) Mallomonas Paraphysomonas Multi-affiliation (Chrysophyceae) Unknown family (Hyphochytriales) Sorodiplophrys Tetramitus Naegleria Petalomonas Unknown genus (Haptoria) Phascolodon Discophrya Unknown genus (Phyllopharyngea) Tetrahymena Multi-affiliation (Oligohymenophorea) Epistylis CED F

Figure 3. Relative abundance and kinetics of abundant bacterial (A) and eukaryotic (B) genera of the 23 time- series SM_WWTP samples. Genera were fltered for those that display an abundance rate ≥ 5% in at least one of the 23 samples. (C–F) Sludge microscopic images illustrating the structure of the SM_WWTP microbiota over the 236 days of time series sampling. (C) Bright-feld microscopy image from day 40 sample (frst phase) × 100. Te image shows bacterial cells starting to organize into colonies and a small a free-living protist swimming beside them. (D) Dark-feld microscopy image from day 230 sample (third phase) × 100. (E) Bright-feld microscopy image from day 40 sample (frst phase) × 400. Te image shows a small free-living protist swimming among dispersed bacterial cells. (F). Dark-feld microscopy image from day 230 sample (third phase) × 100.

Bacterial colonization kinetics at the genus level. Abundant genera kinetics, which appears with at least a 5% abundance rate in at least one of the 23 time-series samples, is presented in Fig. 3A. Tese abundant genera constitute around 60% of the total reads during the frst phase, and decrease to about 25% in the second and third phase. Tis means that the global bacterial diversity increases at the end of the frst phase. Tis increase

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 6 Vol:.(1234567890) www.nature.com/scientificreports/

A 100%

80%

60% Reads 40%

20%

0% 13 16 21 28 30 33 35 40 68 82 133140 147154 182189 196203 210217 224231 236 Days

Apicomplexa Ciliophora DinoflagellataStramenopiles AmoebozoaAncyromo_Apuso_Breviata Discoba Cryptia_Cryptophyta_Cryptomonadales Nucletmycea Haptisia_Haptophyta_Prymnesiophyceae (Haptophyta) Holozoa Metamonada_Preaxostyla Chloroplastidia Rhodophyceae_Porphyridiophyceae Rhizaria_Cercozoa Multi-affiliation unknown phylum BC

100% AlveolataCryptomycota 50% Chytridiomycota LKM15 Stramenopiles Heterolobosea Ichthyosporea 80% 40% Metazoa_Annelida Metazoa_Arthropoda Protosporangiida 60% 30% Reads Reads 40% 20%

20% 10%

0% 0% 3 0 7 4 1 6 3 0 7 4 2 9 3 0 7 4 1 6 6 3 0 7 4 2 9 6 30 33 35 40 68 82 13 16 21 28 DaysD13 16 21 28 30 33 35 40 68 82 ays 19 20 21 21 22 23 23 13 14 14 15 18 18 19 20 21 21 22 23 23 13 14 14 15 18 18 D 100% Euglenozoa 80% Rotifera Rhizaria_Cercozoa 60% Reads 40%

20%

0% 3 0 7 4 2 9 6 0 7 4 1 6 3

13 16 21 28 30 33 35 40 68 82 Days 13 14 14 15 18 18 19 20 21 21 22 23 23

Figure 4. Relative abundance expressed as percent of total 18S rDNA reads, and colonization kinetics of the main eukaryotic phyla over the 236 days of the SM_WWTP colonization. (A) Te relative abundance evolution of the 15 eukaryotic phyla, expressed as percent of total 18S rDNA reads over the 236 days of SM_WWTP colonization. Colored lines delimit the three periods of eukaryotic evolution. Pink represents the frst period, blue, the intermediate period, and green represents the third period. (B) Curves showing the evolution of the frst eukaryotic colonizers, (C) the intermediate eukaryotic colonizers, and (D) the late eukaryotic colonizers.

takes place with a radical change of the genera (Fig. 3A; Tables S2, S3 and S4). Over the frst period (13–40 days), we show an important population of sequence reads, (25–71%), predominantly afliated with, Pseudomon- adales, Rhodobacterales, Caulobacterales, Sphingomonadales, Rhizobiales, and Enterobacteriales, (Fig. 2F). Te four top OTUs over this period afliate with Acinetobacter (Pseudomonadales) 13.4%, Phenylobacterium (Caulo- bacterales) 5.7%, Acidovorax (Burkholderiales) 4.2%, and Stenotrophomonas (Xanthomonadales) 4% (Table S2). Over the second period (40–133 days), two OTUs stand out, afliating with Termomonas (Xanthomonadales) 13.8% and Candidatus Competibacter (Competibacterales) 2.9% (Table S2). Te third period (133–236 days) is characterized by one leading OTU, PLTA13 (Xanthomonadales) 6.5% (Table S4). We notice an evolution of the structure of the wastewater microbial community, from dispersed planktonic bacterial and eukaryotic cells (Fig. 3C,E) to well structured and sessile bacterial and eukaryotic cells within a foc, as it is shown in Fig. 3D,F respectively.

The nitrifcation–denitrifcation process. Physicochemical parameters show that ammonia concentra- tion varied between 53.03 and 55.6 mg L−1 over 40 days, and starts decreasing from day 40 through day 82, end- ing up with 0.23 mg L−1 at day 82. At the same time, we notice an increase of nitrite and nitrate concentrations

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 7 Vol.:(0123456789) www.nature.com/scientificreports/

at day 40 and day 68 respectively (Fig. 1A). Afer that, nitrifcation–denitrifcation processes remain stable over the rest of the time.

Colonization kinetics and phylogenetic diversity of the Eukaryotic population. Implanta- tion kinetics of the major eukaryotic organisms (Fig. 4A) shows three types of populations, the frst coloniz- ers represented by Nucletmycetea (Cryptomycota), Stramenopiles (Ochrophyta), and Alveolata (Apicomplexa, and Ciliophora) (Fig. 4B). Intermediate colonizers, which started colonizing the SM_WWTP afer 40 days, are represented by Nucletmycetea (Chytridiomycetes, LKM15), Heterolobosea, Ichthyosporea, Amoebozoa (Protospo- rangida), Metazoa (Arthropoda, Annelida, Protosporangea), (Fig. 4A,C). Te last group composed of Rhizaria (Cercozoa), Rotifera, and Euglenozoa (Fig. 4D), represents the late colonizers and appears afer day 140. Te same pattern is observed for minor eukaryotic phyla. Kinetics of abundant eukaryotic genera, present at a rate of abundance > 5% at least once in the time series samples, are shown in Fig. 3B. Tey are representing about 85% of the total reads, and remain stable during the time series, in contrast to bacterial genera (Fig. 3A). Tis pattern is not fundamentally diferent from the pat- tern of eukaryotic phyla kinetics (Fig. 4A). Tis may be linked to the fact that phyla are represented mostly by abundant genera. However, there is also a radical change between the frst and the third phase of the time series, as the frst phase genera are replaced by novel abundant genera.

Persistent prokaryotic and eukaryotic populations. Exploration of OTU occurrence shows the per- sistence of 30 and 19 OTUs for both prokaryotic and eukaryotic populations respectively (Fig. S7A, Fig. S7B). Tese OTUs consist of a minimum of one read but display variable abundances throughout the study. Tese persistent OTUs made up merely 0.004% of the total bacterial OTUs; they are afliated with only four phyla: Bacteroidetes, Proteobacteria, Firmicutes and Actinobacteria (Fig. S7A), but accounted for an average of 14% [3–35%] of all 16S rDNA sequence reads. Although, the 19-core eukaryotic OTUs amounted to only 0.013% of the total OTUs, they accounted for an average of 69% [22–98%] of all 18S rDNA reads. Tey are afliated with nine phylogenetic groups: Holozoa (Arthropoda, Rotifera), Nucletmycea (Ascomycota, Cryptomycota), Rhizaria (Cercozoa), Alveolata (Ciliophora), Euglenozoa, Stramenopiles (Ochrophyta, Hyphochytriomycetes, Peronosporo- mycetes), Ichthyosporea, Heterolobosea and unknown phylum (Fig. S7B). Discussion Most ecological studies nowadays describe microbial communities from a perspective of phylogenetic levels, from phyla to genera, or based on species living within complex ecosystems. In microbial ecology, simple observation and description of community’s patterns are not enough, and there is a need to seek the mechanisms that may explain the occurrence of these patterns. A huge gap still exists in understanding how environmental factors and microbe-to-microbe interactions are shaping communities in space and time within complex microbial ecosystems. In the present study, we have tracked microbial species and communities over a period of 236 days in a context where they were sequentially colonizing and building up a stable and efciently functioning ecosystem. With this objective, we have sampled the aerobic basin of the SM_WWTP sequentially from March through October 2014. To obtain a holistic view of the process, we tried to be as complete as possible by analyzing both physicochemical parameters and microbial components (Eukarya and Bacteria) starting with the frst inoculation of the basins with raw sewer microbes through the establishment of a complex, stable and functional microbial community. When they frst arrive in the basin of the SM_WWTP, sewer microbes still diluted in the black water, consti- tute the planktonic phase of the infuent inoculum. Tey start building a robust food web composed of microbial communities well adapted to the available nutrients and the physicochemical parameters of the milieu. Tey start developing multiple physical and metabolic interactions between each other and their habitat, modifying and adjusting these parameters to make and optimize their own living. We cannot explain the colonization kinetics without taking into account the physicochemical parameter evolution and the network of interactions between microbial species over the overall colonization period.

The nitrifcation–denitrifcation process. Classically, in complex and functionally stable wastewater ecosystems, ammonia oxidation is achieved by species afliated with Nitrosomonas europaea, Nitrosomonas eutropha, Nitrosococcus, Nitrosovibrio, Nitrosospira, Nitrosoarchaeum, Nitrosocaldus, Nitrosopumilus, Nitros- osphaera, and Brocadia11,12. However, our study shows that none of these genera was detected over the frst period. Te abrupt decrease in ammonium concentration in the aerobic basin during the intermediate phase (40–33 days) may be explained in part by the metabolic activity of the heterotrophic nitrifers and denitri- fers, and also, by the assimilation of nitrogen by the microbiota, which we have shown to increase and diver- sify over the frst and the intermediate period. Literature examination indicates that microorganisms such as many prokaryotes and fungi carry out heterotrophic nitrifcation and denitrifcation­ 13. Fungi, such as Asper- gillus favus14, Verticillium sp.15, Absidia cylindrospora16, also carry out heterotrophic nitrifcation. It has been demonstrated that isolated fungal strains can remove both nitrogen and phosphorus from ­wastewater17,18. Liu et al., showed that the concentration of ammonia in chicken manure was signifcantly lowered compared to the control group using the fungus Paecilomyces variotii, they isolated from the same ­environment19. By studying competition between the heterotrophic nitrate bacterium Paracoccus denitrifcans and Nitrosomonas europaea (autotrophic bacterium)20, the authors have shown that under limiting oxygenation conditions and for high C/N concentration, bacterial heterotrophic nitrifcation becomes the dominant process and reaches up to 60% of total nitrifcation­ 21. Moreover, numerous studies show that denitrifcation is usually carried out by faculta- tive anaerobes such as gram-negative classes of α-, β- and γ-Proteobacteria (e.g. Acinetobacter calcoaceticus,

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 8 Vol:.(1234567890) www.nature.com/scientificreports/

Alcaligenes faecalis, Microvirgula aerodenitrifcans, Paracoccus denitrifcans, Tiobacillus denitrifcans, and the genera Acidovorax, Agrobacterium, Azoarcus, Tauera, Hyphomicrobium, Methylobacterium, Rhodobacter, or the families Pseudomonadaceae, Comamonadaceae, and Rhodocyclaceae22–29. Meanwhile, many diazotrophic micro- organisms, which fx atmospheric nitrogen gas into a more usable form, such as some species of Azospirillum and Bradyrhizobium, are able to ­denitrify30. Denitrifcation is also found among a few Archaea31,32 and Fungi33,34, including Ascomycota (e.g., Fusarium oxysporum, Fusarium solani, Cylindrocarpon tonkinense and Gibberella fujiuroii) and Basidiomycota (e.g., Trichosporon cutaneum). Te fungi Fusarium oxysporum and Cylindrocar- pon tonkinense constitute veritable denitrifers. Tis denitrifcation was confrmed by observations showing that nitrate and nitrite reductions occurs in the ­mitochondria35. On the other hand, it has been demonstrated that these operations are coupled to a net synthesis of ATP­ 36. Denitrifcation has also been described in some ­foraminifera37. In the present study, sequences afliated with above orders, families or genera involved in heterotrophic nitrifcation–denitrifcation are present at high abundance within the frst phase of the time series samples. Tey accounted for between 19 and 95% of the total sequence reads during the frst phase and only between 6 and 14% over the second phase. We think that, the nitrifcation–denitrifcation process may be predominantly triggered by heterotrophic prokaryote and eukaryote components of the ecosystem over the frst period. During the intermediate and the third phase (40–236 days), we show a gradual appearance of functional groups not present before, or present, but at a low level, such as, the nitrifying-denitrifying lithoautotrophic bacteria, hydrolyzing bacteria, and phosphate accumulating bacteria. Autotrophic nitrifying bacteria are usually associated with the focs as expected from these organisms that grow in tight micro-colonies embedded in sludge focs in order to perform their specifc function, joined by the denitrifers in tight ­association38. Ammonia oxidiz- ing bacteria afliated with either Nitrosospira or Nitrosomonas genera and some gram-positive bacteria such as Bacillus licheniformis, Paracoccus denitrifcans have been shown to be capable of denitrifcation­ 39–43. Candidatus Accumulibacter is a genus afliating with polyphosphate-accumulating organisms which uses nitrite as electron 44 acceptor for denitrifying phosphorus ­removal . Tis fnding is in line with recent reports showing that ­N2 pro- duction by denitrifcation is higher in particulate matter than in the planktonic phase in both marine and river environments, correlating with the organic content of the ­particles45. In our study, the autotrophic nitrifying- denitrifying bacteria were undetectable during the frst planktonic phase, where aggregates or focs were still under formation. Te autotrophic ammonia and nitric oxidizers appear starting from day 82 (Figs. 1A, 2E).

Exopolysaccharide secretion and focs formation. Microorganisms take advantage of the nutriment- rich wastewater to synthetize reserve substances, among which, the exopolysaccharides (EPS) and organize species-rich structures in the form of microbial aggregates and planktonic bioflms called focs. In our study, we observe that microbial populations are gradually switching from mostly planktonic to a more aggregated sessile life. Te appearance of well-structured focs requires between 2 and 3 months of adaptation and acclimation to the novel habitat, and provides stability and durability of metabolic interaction within the ecosystem. Hence, microscopy images presented in both, Fig. 3D,F show well-structured microbial cells, where bacteria and ses- sile protozoa (probably Vorticella), are tightly attached within the foc matrix. Our fnding is in good agreement with studies showing a constant enrichment of EPS producing microorganisms over the frst period. Tis period favors biomass structuration that supports process efciency and stability. Microbial EPS are an abundant and important group of compounds that can be secreted by Archaea, Bacteria, Fungi and Algae­ 46–48. EPS protect bacteria from environmental stresses, tolerate higher concentrations of many biocides, and play a defnite role in sludge focculation, that helps microorganisms efciently biodegrades a wider range of substrates than pure cul- tures of free-living microorganisms. Within EPS, microorganisms can establish stable arrangements and func- tion as synergistic micro-consortia such is shown in Fig. 3D,F, enabling the cells to function in a manner similar to multicellular organisms and complement each other’s functions. Te EPS matrix facilitates nutrient seques- tration, the retention of exo-enzymes, cellular debris and genetic material. It can be considered as a microbial recycling ­yard46. Moreover, EPSs are of considerable importance in the removal of pollutants from wastewater, in bio-focculation and settling and in the sludge dewatering. Bacteria afliated with orders and families of Firmicutes (Leuconostocaceae and Streptococcaceae), and Proteobacteria (Burkholderiales, Pseudomonadales and Xanthomonadales), among others have been described as producers of ­polysaccharides49,50. Within the transitional period (40–133 days), Acinetobacter represents only 2.2% of dominant genera; it is gradually replaced by novel OTUs afliated namely with Termomonas, 13.8%, Candidatus Competibacter 2.9%, and Arenimonas 2%, etc. (Table S2). Tis intermediate phase composed of 371 persistent OTUs, making up to 76% of the total reads, with a predominant bacterial composition completely diferent from the frst period or the intermediate period. Te third phase is marked by the settlement of novel functionally major colonizers such as Planctomycetes, Acidobacteria, Patescibacteria, Chlorofexi, Gemmatimonadetes, Nitrospirae, Hydrogenedentes, Armatimonadetes, and WPS2 (Fig. 2D,E), which may explain the high values of Shannon, Inverse Simpson diversity indices and evenness. More importantly, the fve top abundant OTUs are afliated mainly with Xanthomonadales, Chitin- ophagales and Chlorofexi: unknown OTU PLTA13 (6.5%), an unknown OTU (2.7%), the genera Dokdonella (2.3%), Terrimons (2.1%), and unknown Chlorofexi (Table S4). Xanthomonadales and Chitinophagales are known to be EPS producers­ 51,52. Members of the Chitinophagaceae family dominating the sludge during the third period, are aerobic het- erotrophs capable of degrading organic matter and are involved in biosynthesizing and exporting extracellular polymeric substances (EPSs)51. Among the Chitinophagales, Ferruginibacter is among the top 10 bacterial pre- dominant OTUs; represents 1.5% of the total reads within the third period of colonization of the SM_WWTP. Previous studies reported that Ferruginibacter are highly enriched in sludge and potentially associated with

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 9 Vol.:(0123456789) www.nature.com/scientificreports/

bio-focculation and bioflms­ 53,54 consistently supporting their suggested role in foc formation. In a recent report, Saunders et al., analyzed the microbial communities in 13 Danish wastewater treatment plants in consecutive years and a single plant periodically over 6 years, using Illumina sequencing of 16S ribosomal RNA amplicons of the V4 ­region4. Te plants contained a core community of 63 abundant genus-level OTUs that made up 68% of the total reads. Tey showed that Chitinophagaceae and Comamonadaceae represent a core bacterial members of sludge that frequently occur across many full-scale, geographically diferentially located WWTPs. Hence, this study shows that structural and metabolic complexity of the ecosystem involves EPS synthetizing bacteria, which predominate over the second and the third period of the colonization process. Our results show the existence of high proportion of longstanding prokaryotic and mostly eukaryotic per- sistent species in activated sludge. Tey may sustain the long-term ecosystem functional stability. While the abundances of bacterial OTUs decreases steadily throughout the period of sampling (Fig. S7A), the eukaryotic persistent OTUs maintain their abundances within a frame of 60–100% of the total eukaryotic reads (Fig. S7B). In summary, we could distinguish clearly 2–3 types of population patterns within both Bacteria and Eukarya domains at various phylogenetic levels (Phylum, Order or OTU): frst, r-strategists referred to as “opportunistic”, or frst colonizers, are free living microorganisms, emphasize high growth rates with less diversity, typically exploiting less-crowded ecological niches and producing many ofspring. Second, K-Strategists or late coloniz- ers, are selected species described as “equilibrium”. Tey display traits mostly associated within focs, close to carrying capacity of the aerobic basin and typically are strong competitors in such crowded niches. Finally, the continuous spectrum populations, which are populations withstanding the physicochemical transformations occurring in the ecosystem. Tey represent the persistent OTUs all over the colonization period. Conclusion and perspectives Physicochemical and microbial parameters show that the wastewater microbiome is established sequentially over a long period. Tey start in part by heterotrophic nitrifcation–denitrifcation processes and become auto- and heterotrophic in the second phase when the focs are well structured. Te development of highly struc- tured focs triggered by EPS producing microorganisms, which become predominant with time, creates specifc anaerobic niches triggering among other functions, the autotrophic nitrifcation–denitrifcation processes. Te assemblage of various microorganisms within focs favors multitude of metabolic interactions, which may be assessed by multi-meta-omics (meta-transcriptomics, metabolomics and meta-proteomics) technologies. Whole shotgun sequencing metagenomes of representative samples from each phase of colonization represents a frst step toward understanding how the microbiota is gradually creating its living and the pattern of the main func- tions established over time toward the building of functionally stable ecosystem. Importantly, the onset of the nitrifcation process is correlated with radical changes in both prokaryotic and eukaryotic populations, leading from a planktonic mode of growth of microbial populations, to a fully functional wastewater treatment plant, associated with populations organized around foc structures. Further characterization of the microbiota by sequencing the WWTP viriome is more than needed to have a holistic view of the microbial interactions and to explore the networks of relationships between the continu- ously evolving microbial community members of the SM_WWTP microbiome. Based on their presence, their richness, the abiotic factors (physicochemical parameters), and inter-taxa correlations, taxon co-occurrence or exclusion patterns, we will be able to gain more integrated understanding of microbial communities’ structure and the ecological rules guiding assembly of complex microbial communities across temporal gradients. Material and methods Wastewater treatment plant description and sampling. Te SM_WWTP, is located at Blanc-Mesnil (48° 57′ 09.6″ N 2° 27′ 46.5″ E, Seine-Saint-Denis, France), was commissioned in March 2014. Te SM_WWTP has treatment capacity of 50,000-m3 day­ −1, which can be extended to 76,500-m3 day−1 in rainy weather. It treats wastewater from a residential area of 200,000 inhabitants in the northeastern of Paris and efuents from Roissy- Charles de Gaulle International Airport. It discharges its treated efuents into the Morée stream, a 12.4 km long river, fowing into the Croult, which joins the Seine River. Tis river has been heavily impacted by the lack of a wastewater treatment plant since the nineteenth century. Te purpose/goal of the SM-WWTP is to restore the quality of its water in accordance with the European Water Framework Directive. Tis plant started de novo without any external sludge inoculation. It was flled up with potable water on March 3rd, 2014, and gradually supplied with raw wastewater. Te plant performance data was made available by the wastewater facility. Water temperature, Biochemical oxygen demand (BOD), Chemical oxygen demand (COD), pH, gross fow rate discharge, total suspended solids (TSS), oxygen concentration in aerobic sludge, set- tling volume, dryness, total volatile suspended solids, Mohlman index, nitrogen (total Kjeldahl nitrogen (TKN), ammonia (N-NH4), nitrite (N-NO2 and nitrate N-NO3) and phosphorus concentrations were determined accord- ing to standard methods. Physicochemical parameters measurements were performed every day by the SIAAP (Syndicat Interdépartemental pour l’Assainissement de l’Agglomération Parisienne) laboratory at the entrance and the exit of the biological tank. Te measurements were performed according to the French standards methods: TSS, according to NF EN 872; TKN, according to NF EN 25663; Ammonia, according to NF EN ISO 11732; nitrite and nitrate, according to NF EN ISO 13395; Orthophosphate and total P, according to NF EN ISO 6878. For the sequencing work, the sampling was done independently of physicochemical parameter measurements. A summary of important plant chemical and operational parameters is shown in Table S1. Activated sludge samples were collected from the aerobic basin for 237 consecutive days (starting from March 18th, 2014 to October 27th). For each sampling point, 2 L of wastewater samples were directly collected at the end of the aeration tank. Collected wastewater samples were shipped refrigerated at 4 °C and analyzed within

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 10 Vol:.(1234567890) www.nature.com/scientificreports/

6 h. Samples were then concentrated by centrifugation at 6,500g for 15 min at 4 °C. Te pellets of the wastewater samples were kept at − 20 °C prior to metagenomic DNA extraction.

DNA extraction, amplifcation and sequencing. DNA extraction was performed using Nucleo spin soil DNA kit (Macherey Nagel GmbH & Co. KG, Durën, Germany). DNA extracts were quantifed by a spec- trophotometric method using the WPA Biowave II UV/Visible spectrophotometer (Biochrom, Cambridge, UK) and a TrayCell Fibre optic micro cell (Hellma GmBH & Co. KG, Müllheim, Germany). Prokaryotic diversity was investigated by PCR amplifcation of the hypervariable region V4-V5 of the SSU rDNA gene with fusion primers 515F (5′-Ion adapter–Barcode–GTG​YCA​GCMGCC​GCG​GTA-3′) and 928R (5′-Ion trP1 adapter–CCCCGY​ CAA​ TTC​ MTTT​ RAG​ T-3​ ′)55, which include a barcode and sequencing adapters. Te fusion PCR method uses fusion primers to fx the Ion A adapter (5′-CCATCT​ CAT​ CCC​ TGC​ GTG​ TCT​ CCG​ ​ ACT​CAG​-3′) linked to a barcode, and the Ion truncated P1 (trP1) adapter (5′-CCT​CTC​TAT​GGG​CAG​TCG​ GTGAT-3′) to the amplicons as they are generated during PCR. Te PCR mix contained 1X Pfx amplifcation Bufer, 0.3 mM of each dNTP, 1 mM MgSO­ 4, 0.3 µM of each primer, 1U Platinum Pfx DNA polymerase (Invitro- gen), and 10–20 ng of template DNA in a 50 µL reaction volume. Amplifcation was performed as follows: 5 min at 94 °C, 30 cycles of 15 s at 94 °C, 30 s at 50 °C, 1 min at 68 °C, followed by fnal extension of 5 min at 68 °C. PCR products were purifed using Agencourt AMPure XP magnetic beads (Beckman Coulter) according to the manufacturer’s instructions, with a bead versus amplicon ratio of 1.2, and eluted in 45 µL TE Bufer (10 mM Tris–HCl pH 8.0, 1 mM EDTA). Purifed amplicons were quantifed using DNA 1,000 Kit and 2,100 Bioanalyzer (Agilent Technologies), following the manufacturer’s instructions. Ten, all amplicons were pre-diluted at 500 pM in molecular grade water and equimolarly pooled. Te pool was then diluted at 100 pM for sequencing. Briefy, to prepare template-positive Ion Sphere Particles (ISPs) containing clonally amplifed DNA by emulsion PCR, the library was diluted to 26 pM and set up on the Ion OneTouch 2 Instrument (Life Technologies) using the Ion PGM Hi-Q View OT2 Kit (Life Technologies) following the manufacturer’s instructions. Tese templated ISPs were purifed on the Ion OneTouch ES (Life Technologies) according to the manufacturer’s instructions. Prokaryotic rDNA gene sequencing was performed on Ion Torrent Personal Genome Machine using Ion 316 Chip V2 (Life Technologies) and the Ion PGM Hi-Q View Sequencing Kit (Life Technologies) according to the manufacturer’s instructions. Te Torrent suite sofware processed sequencing data. Te sofware fltered out low quality and polyclonal sequence reads, and quality fltered data were exported as FastQ fles. Eukaryotic diversity was investigated by PCR amplifcation of the V9 hypervariable region of the SSU rRNA genes using the two primers 1389F/1510R (25 cycles in triplicate). Te resulting 150 bp PCR fragments were subjected to Illumina libraries preparations that were sequenced on MiSeq platform at Genoscope (Evry, France). Although a comparison of diferent 18S rRNA gene targeting primers within the V4 region performed ­best56 we used the V9 barcode for the following reasons. (1) it presents a combination of advantages for addressing general questions of eukaryotic biodiversity over extensive taxonomic and ecological scales, (2) it is universally conserved in length (130 ± 4 bp) and simple in secondary structure, thus allowing relatively unbiased PCR amplifcation across eukaryotic lineages followed by Illumina sequencing, (3) it includes both stable and highly- variable nucleotide positions over evolutionary time frames, allowing discrimination of taxa over a signifcant phylogenetic depth, and (4) it is extensively represented in public reference databases across the eukaryotic tree of life, allowing taxonomic assignment amongst all known eukaryotic ­lineages57. Despite the high rarefaction coverage and the presence of sequences belonging to many major micro-eukaryotic taxa, we cannot exclude that our primer set covering the V9 region of the 18S rRNA gene may have not sufciently mapped the diversity in particular of non-ciliate groups or minor populations in the SM_WWTPs.

Sequence quality control and bioinformatics processing. For 18S rDNA Illumina reads, quality control began by removing adapters and primers on the whole reads and low quality nucleotides from both ends, and then we continued the next steps using the longest sequence without adapters or low quality bases. Reads shorter than 30 nucleotides afer trimming and read pairs that come from the low-concentration spike-in library of Illumina PhiX Control were discarded. Tis policy allows submission of high quality data (without contami- nation) in order to interrogate databases and to improve subsequent analysis. Overlapping 18S rDNA paired end reads were merged with pear v0.9.11 (https​://githu​b.com/easyb​uilde​rs/easyb​uild-easyc​onfg​s/pull/6653/ fles​). For 16S rDNA reads, we used cutadapt 1.12 (https​://genso​f.paste​ur.fr/docs/cutad​apt/1.12/guide​.html) to remove adapters, primers, and discarded reads with low quality nucleotides (when quality value is < 20). Dereplicated 18S and 16S rDNA reads were independently clustered with swarm 2.1.12 (https://biowe​ b.paste​ ​ ur.fr/packages/pack@swarm​ @2.1.12​ ), using a distance cutof of 3%, and singleton OTUs were removed. Chimeric sequences were detected with VSEARCH (https​://githu​b.com/torog​nes/vsear​ch), and removed for subsequent analyses. 16S and 18S rDNA sequence analyses were continued using the FROGS pipeline “Find, Rapidly, OTUs with Galaxy Solution” (https​://frogs​.toulo​use.inrae​.fr/). Taxonomic afliation of 16S and 18S rDNA reads was per- formed with BLAST 2.6 on SILVA_132_16S and SILVA_132_18S databases respectively. A Biological Observation Matrix fle (BIOM) comprising both abundance and taxonomy, was generated and imported into R (version 3.5.2) for statistical analysis.

Statistical analyses. Alpha diversity estimators were computed using the Phyloseq sofware package­ 58, which is an integrated algorithm into FROGS pipeline. Briefy, calculation of community richness and alpha- diversity indices (e.g. Shannon–Weiner and InvSimpson) was done using BIOM fles. In this paper, we show only the evolution of Shannon–Weiner and inverse Simpson indices over time. A hierarchically clustered heat-map was generated using the top 100 microbial OTUs respectively for prokaryotes and eukaryotes respectively from

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 11 Vol.:(0123456789) www.nature.com/scientificreports/

the 23 time-series SM_WWTP samples. We used Bray–Curtis dissimilarity to quantify the diferences between OTUs and species among the time series SM_WWTP samples.

Principal component and canonical component analyses. In order to get insights into the relation- ships between WWTP time series samples, centered, scaled, principal component analyses (PCoA) were imple- mented on 16S rDNA and 18S rDNA sequence tags datasets respectively using ADE4 R ­package59. Considering the dispersion in the total number of OTUs reads identifed in each sample, OTUs abundances were scaled by dividing the number of reads of each OTU in a given sample by the sum of total reads for the same sample. We considered only OTUs that exceeded 1% in at least one sample for the analysis. We kept 117 OTUs of 16S rDNA and 98 OTUs of 18S rDNA for subsequent analyses. We took into account OTUs variable distributions represented by axes accounting for the highest percent of the total variance. To get insight into the relation- ships between physicochemical and bacterial communities evolution, we implemented canonical correspond- ence Analysis (CCA) using vegan R package. PCoA and CCA. For both PCoA and CCA, fgures were visualized using ggplot2.

Originality‑signifcance‑statement. Te originality of the study resides in its 8-month sequential follow- up of both bacterial and micro-eukaryotic community colonization of a newly constructed wastewater treatment plant (WWTP) seeded by raw water microorganisms rather than sludge from a diferent WWTP as is commonly done. To the best of our knowledge, we do not know of any study that analyzes microbial colonization kinetics from the beginning of operations up to the establishment of a complex environmental microbial ecosystem. It leads to a holistic view of the wastewater microbiome structure, population dynamics and the kinetics of its establishment, thanks to high throughput sequencing technology and sampling of a newly constructed WWTP over time. We were able to distinguish mainly three populations patterns. Te frst colonizers, represented by unstructured populations with high growth rates, typically exploit less-crowded ecological niches and produce many ofspring; the second colonizers, that take over from the frst colonizers form a structured population, we call K-Strategists or equilibrium strategists. Tey display traits associated with living at densities close to the carrying capacity of the milieu; and fnally, continuous spectrum populations, that may represent the backbone of microbial communities. Tis group shows strong traits of longevity and may guarantee microbial community functioning, due to the bufering capacity they provide to the ecosystem to help resist environmental change. Data availability Raw data for relative abundance of both eukaryotic and bacterial communities at the diferent taxonomic lev- els will be made available and provided on reasonable request. Sequences reported in this study were depos- ited in EMBL databases (https​://www.ebi.ac.uk/) under accession numbers ERX4094207-ERX4094229; and ERR4106944-ERR4106967 for eukaryotic sequences and ERS4556742 to ERS4556765 for bacterial sequences.

Received: 28 January 2020; Accepted: 24 June 2020

References 1. Grady, C. P. L., Daigger, G. T. & Love, N. G. Biological Wastewater Treatment (IWA Publishing, London, 2011). 2. Wagner, M. & Loy, A. Bacterial community composition and function in sewage treatment systems. Curr. Opin. Biotechnol. 13, 218–227 (2002). 3. Ju, F. & Zhang, T. Bacterial assembly and temporal dynamics in activated sludge of a full-scale municipal wastewater treatment plant. Isme J. 9, 683–695 (2015). 4. Saunders, A. M., Albertsen, M., Vollertsen, J. & Nielsen, P. H. Te activated sludge ecosystem contains a core community of abundant organisms. Isme J. 10, 11–20 (2016). 5. Matsunaga, K., Kubota, K. & Harada, H. Molecular diversity of eukaryotes in municipal wastewater treatment processes as revealed by 18S rRNA gene analysis. Microbes Environ. 29, 401–407 (2014). 6. Chouari, R. et al. Eukaryotic molecular diversity at diferent steps of the wastewater treatment plant process reveals more phylo- genetic novel lineages. World J. Microbiol. Biotechnol. 33, 017–2217 (2017). 7. Matsubayashi, M., Shimada, Y., Li, Y. Y., Harada, H. & Kubota, K. Phylogenetic diversity and in situ detection of eukaryotes in anaerobic sludge digesters. PLoS ONE 12, e0172888 (2017). 8. Cohen, Y. et al. Bacteria and microeukaryotes are diferentially segregated in sympatric wastewater microhabitats. Environ. Micro- biol. 21, 1757–1770. https​://doi.org/10.1111/1462-2920.14548​ (2019). 9. Flowers, J. J., Cadkin, T. A. & McMahon, K. D. Seasonal bacterial community dynamics in a full-scale enhanced biological phos- phorus removal plant. Water Res. 47, 7019–7031 (2013). 10. Zhang, B. et al. Seasonal bacterial community succession in four typical wastewater treatment plants: correlations between core microbes and process performance. Sci. Rep. 8, 4566. https​://doi.org/10.1038/s4159​8-018-22683​-1 (2018). 11. Stein, L. Y., Arp, D. J. & Hyman, M. R. Regulation of the synthesis and activity of ammonia monooxygenase in Nitrosomonas europaea by altering pH to afect NH(inf3) availability. Appl. Environ. Microbiol. 63, 4588–4592 (1997). 12. Chain, P. et al. Complete genome sequence of the ammonia-oxidizing bacterium and obligate chemolithoautotroph Nitrosomonas europaea. J. Bacteriol. 185, 2759–2773 (2003). 13. Pedersen, H., Dunkin, K. A. & Firestone, M. K. Te relative importance of autotrophic and heterotrophic nitrifcation in a conifer forest soil as measured by 15N tracer and pool dilution techniques. Biogeochemistry 44, 135–150 (1999). 14. White, J. P. & Johnson, G. T. Afatoxin production correlated with nitrifcation in Aspergillus favus group species. Mycologia 74, 718–723 (1982). 15. Lang, E. Fungi of a forest soil nitrifying at low pH values. FEMS Microbiol. Ecol. 38, 257–265 (1986). 16. Stroo, H. F. Heterotrophic nitrifcation in an acid forest soil and by an acid-tolerant fungus. Appl. Environ. Microbiol. 52, 1107–1111 (1986). 17. Guest, R. & Smith, D. Isolation and screening of fungi to determine potential for ammonia nitrogen treatment in wastewater. J. Environ. Eng. Sci. 6, 209–217 (2007).

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 12 Vol:.(1234567890) www.nature.com/scientificreports/

18. Hultberg, M. & Bodin, H. Fungi-based treatment of brewery wastewater—biomass production and nutrient reduction. Appl. Microbiol. Biotechnol. 101, 4791–4798 (2017). 19. Liu, Z. et al. Paecilomyces variotii: a fungus capable of removing ammonia nitrogen and inhibiting ammonia emission from manure. PLoS ONE 11, e0158089 (2016). 20. Moir, J. W., Crossman, L. C., Spiro, S. & Richardson, D. J. Te purifcation of ammonia monooxygenase from Paracoccus denitri- fcans. FEBS Lett. 387, 71–74 (1996). 21. Jetten, M. S. et al. Novel principles in the microbial conversion of nitrogen compounds. Antonie Van Leeuwenhoek 71, 75–93 (1997). 22. Beller, H. R. et al. Te genome sequence of the obligately chemolithoautotrophic, facultatively anaerobic bacterium Tiobacillus denitrifcans. J. Bacteriol. 188, 1473–1488 (2006). 23. Joo, H. S., Hirai, M. & Shoda, M. Improvement in ammonium removal efciency in wastewater treatment by mixed culture of Alcaligenes faecalis no. 4 and L1. J. Biosci. Bioeng. 103, 66–73 (2007). 24. Su, J. J., Liu, B. Y. & Liu, C. Y. Comparison of aerobic denitrifcation under high oxygen atmosphere by Tiosphaera pantotropha ATCC 35512 and Pseudomonas stutzeri SU2 newly isolated from the activated sludge of a piggery wastewater treatment system. J. Appl. Microbiol. 90, 457–462 (2001). 25. Patureau, D. et al. Combined phosphate and nitrogen removal in a sequencing batch reactor using the aerobic denitrifer, Micro- virgula aerodenitrifcans. Water Res. 35, 189–197 (2001). 26. Kim, M. et al. Aerobic denitrifcation of Pseudomonas putida AD-21 at diferent C/N ratios. J. Biosci. Bioeng. 106, 498–502 (2008). 27. Zhao, B., He, Y. L., Hughes, J. & Zhang, X. F. Heterotrophic nitrogen removal by a newly isolated Acinetobacter calcoaceticus HNR. Bioresour. Technol. 101, 5194–5200 (2010). 28. Willems, A. Te family Comamonadaceae. In Te Prokaryotes: Alphaproteobacteria and Betaproteobacteria 4th edn (eds Rosenberg, E. et al.) 777–851 (Springer, Berlin, 2014). 29. Wu, Y., Shukal, S., Mukherjee, M. & Cao, B. Involvement in denitrifcation is benefcial to the bioflm lifestyle of Comamonas testosteroni: a mechanistic study and its environmental implications. Environ. Sci. Technol. 49, 11551–11559 (2015). 30. Rosch, C., Mergel, A. & Bothe, H. Biodiversity of denitrifying and dinitrogen-fxing bacteria in an acid forest soil. Appl. Environ. Microbiol. 68, 3818–3829 (2002). 31. Bartossek, R., Nicol, G. W., Lanzen, A., Klenk, H. P. & Schleper, C. Homologues of nitrite reductases in ammonia-oxidizing archaea: diversity and genomic context. Environ. Microbiol. 12, 1075–1088 (2010). 32. Cabello, P., Roldan, M. D. & Moreno-Vivian, C. Nitrate reduction and the nitrogen cycle in archaea. Microbiology 150, 3527–3546 (2004). 33. Shoun, H., Kim, D. H., Uchiyama, H. & Sugiyama, J. Denitrifcation by fungi. FEMS Microbiol. Lett. 73, 277–281 (1992). 34. Hayatsu, M., Tago, K. & Saito, M. Various players in the nitrogen cycle: Diversity and functions of the microorganisms involved in nitrifcation and denitrifcation. Soil Sci. Plant Nutr. 54, 33–45 (2008). 35. Pelmont, J. Biodégradations et métabolismes: les bactéries pour les technologies de l’environnement (ed. Les Ulis EDP Sciences DL) (Coll. Grenoble Sciences, 2005). 36. Kobayashi, M. et al. Denitrifcation, a novel type of respiratory metabolism in fungal mitochondrion. J. Biol. Chem. 271, 16263– 16267 (1996). 37. Glud, R. N. et al. Nitrogen cycling in a deep ocean margin sediment (Sagami Bay, Japan). Limnol. Oceanogr. 54, 723–734 (2009). 38. Dolinsek, J., Lagkouvardos, I., Wanek, W., Wagner, M. & Daims, H. Interactions of nitrifying bacteria and heterotrophs: identifca- tion of a Micavibrio-like putative predator of Nitrospira spp. Appl. Environ. Microbiol. 79, 2027–2037 (2013). 39. Shaw, L. J. et al. Nitrosospira spp. can produce nitrous oxide via a nitrifer denitrifcation pathway. Environ. Microbiol. 8, 214–222 (2006). 40. Medhi, K., Singhal, A., Chauhan, D. K. & Takur, I. S. Investigating the nitrifcation and denitrifcation kinetics under aerobic and anaerobic conditions by Paracoccus denitrifcans ISTOD1. Bioresour. Technol. 242, 334–343 (2017). 41. Khanichaidecha, W., Ratananikom, A. N. K., Eamrat, R. & Kazama, F. Heterotrophic nitrifcation and aerobic denitrifcation using pure-culture bacteria for wastewater treatment. J. Water Reuse Desalin. 9, 10–17 (2019). 42. Amann, R. I., Ludwig, W. & Schleifer, K. H. Phylogenetic identifcation and in situ detection of individual microbial cells without cultivation. Microbiol. Rev. 59, 143–169 (1995). 43. Khanichaidecha, W., Nakaruk, A., Ratananikom, K., Eamrat, R. & Kazama, F. Heterotrophic nitrifcation and aerobic denitrifca- tion using pure-culture bacteria for wastewater treatment. J. Water Reuse Desalin. 9, 10–17 (2019). 44. Zeng, W., Li, B., Wang, X., Bai, X. & Peng, Y. Infuence of nitrite accumulation on “Candidatus Accumulibacter” population structure and enhanced biological phosphorus removal from municipal wastewater. Chemosphere 144, 1018–1025 (2016). 45. Ganesh, S. et al. Size-fraction partitioning of community gene transcription and nitrogen metabolism in a marine oxygen minimum zone. Isme J. 9, 2682–2696 (2015). 46. Flemming, H. C. & Wingender, J. Relevance of microbial extracellular polymeric substances (EPSs)—part i: structural and ecologi- cal aspects. Water Sci. Technol. 43, 1–8 (2001). 47. Wingender, J., Tomas, R. N. & Flemming, H.-C. Microbial Extracellular Polymeric Substances: Characterization, Structure and Function (Springer, Berlin, 1999). 48. More, T. T., Yadav, J. S., Yan, S., Tyagi, R. D. & Surampalli, R. Y. Extracellular polymeric substances of bacteria and their potential environmental applications. J. Environ. Manag. 144, 1–25 (2014). 49. Naessens, M., Cerdobbel, A., Soetaert, W. & Vandamme, E. J. Leuconostoc dextransucrase and dextran: production, properties and applications. J. Chem. Technol. Biotechnol. 80, 845–860 (2005). 50. Rehm, B. Microbial Production of Biopolymers and Polymer Precursors: Applications and Perspectives (Caister Academic, Poole, 2009). 51. Liu, T., Mao, Y. J., Shi, Y. P. & Quan, X. Start-up and bacterial community compositions of partial nitrifcation in moving bed bioflm reactor. Appl. Microbiol. Biotechnol. 101, 2563–2574 (2017). 52. Sutherland, I. W. Microbial polysaccharides from Gram-negative bacteria. Int. Dairy J. 11, 663–674 (2001). 53. Wang, X., Hu, M., Xia, Y., Wen, X. & Ding, K. Pyrosequencing analysis of bacterial diversity in 14 wastewater treatment systems in China. Appl. Environ. Microbiol. 78, 7042–7047 (2012). 54. Faust, L. et al. Characterization of the bacterial community involved in the biofocculation process of wastewater organic matter in high-loaded MBRs. Appl. Microbiol. Biotechnol. 99, 5327–5337. https​://doi.org/10.1007/s0025​3-015-6402-y (2015). 55. Wang, Y. & Qian, P. Y. Conservative fragments in bacterial 16S rRNA genes and primer design for 16S ribosomal DNA amplicons in metagenomic studies. PLoS ONE 4, 0007401 (2009). 56. Hugerth, L. W. et al. Systematic design of 18S rRNA gene primers for determining eukaryotic diversity in microbial consortia. PLoS ONE 9, e95567 (2014). 57. De Vargas, C. et al. Eukaryotic plankton diversity in the sunlit ocean. Science 348, 874 (2015). 58. McMurdie, P. J., Holmes, S. & Watson, M. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE 8, e6127 (2013). 59. Tioulouse, J. Statistical analysis and graphical display of multivariate data on the Macintosh. Comput. Appl. Biosci. 5, 287–292 (1989).

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 13 Vol.:(0123456789) www.nature.com/scientificreports/

Acknowledgements Te authors are very grateful to Suzanne Cure and Adnane Boualem for reading the manuscript, and Olivier Chapleur for his help with bioinformatics analyzes. Authors would like to thank the Genoscope, and INRAE PROSE sequencing teams for providing us with the sequencing data, and especially to Shahinaz Gas and Fred- erick Gavory for their technical support. Author contributions L.M., J.-J.P., J.-L.A., and A.S., wrote the manuscript text. A.G., K.P., and P.B., performed the molecular biology work and the 16S rDNA sequencing. C.M., performed the statistical analyses. K.L., A.L., performed the 18S rDNA sequencing. O.M., L.A., and J.-L.A., provided the samples and carried out the physicochemical analyses.

Competing interests Te authors declare no competing interests. Additional information Supplementary information is available for this paper at https​://doi.org/10.1038/s4159​8-020-68496​-z. Correspondence and requests for materials should be addressed to A.S. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access Tis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/.

© Te Author(s) 2020

Scientific Reports | (2020) 10:11634 | https://doi.org/10.1038/s41598-020-68496-z 14 Vol:.(1234567890)